| 학회 |
한국화학공학회 |
| 학술대회 |
2003년 봄 (04/25 ~ 04/26, 순천대학교) |
| 권호 |
9권 1호, p.198 |
| 발표분야 |
공정시스템 |
| 제목 |
Design of a Wiener-type Input/output Data-based Predictive control without iteration in identification procedure |
| 초록 |
A one step identification method of the Wiener model and the design of a Wiener-type input/output data-based predictive controller on the basis of the identified model are proposed. The linear time invariant and inverse static nonlinear parts can be simultaneously identified without iteration by adopting the subspace identification method. The matrix input/output equation is reformulated in such a way that the output of the nonlinear process can be directly inserted in the matrix input/output equation. Actual calculation of respective identified models is carried out via QR decomposition of the modified block Hankel matrices of the past data. Through the simulation study for the control of multivariate polymerization reactor, this identification method and the performance of the designed predictive controller are examined. The Wiener model identified without iteration turns out to be in good agreement with the first principles model of a solution polymerization reactor and the controller shows satisfactory performances for both set-point tracking and disturbance rejection in the polymer quality control. |
| 저자 |
송인협, 이현구
|
| 소속 |
서울대 |
| 키워드 |
MPC; Wiener model; Identification; Inverse identification
|
| E-Mail |
, |
